Document Type

Thesis - University Access Only

Award Date


Degree Name

Master of Science (MS)

Department / School



Remote sensing is an important tool for land cover studies. Recent advancements in sensor technology have provided the scientific community with a new generation of satellite sensors. Part of the advancement in technology is that sensor bandwidths have been manipulated to include or exclude specific areas of the electromagnetic spectrum. Three sensors with daily global coverage are currently available to scientists. These are the Moderate Resolution Imaging Spectroradiometer (MODIS), Systeme pour l'Observation de la Terre (SPOT) VEGETATION, and Advanced Very High Resolution Radiometer (AVHRR). Differences in spectral bandwidth and other technologies between these sensors raise concerns over the ability to incorporate data from two or more of the sensors into one vegetation study. The Normalized Difference Vegetation Index (NOVI) is often used in vegetation studies to measure the amount of green vegetation present. Vegetation is unique among land cover types in that it absorbs red light and reflects near infrared light. These two properties can be used to determine the amount of vegetation present through the use of the equation NDVI = ((NIRred)/(NIR+red)) This thesis establishes the relationship of NDVI between MODIS, SPOT VEGETATION, and AVHRR over coniferous forest, deciduous forest, grassland, and corn/soybean agriculture. It was found that agreement between the sensors was dependent on the time of year and both vegetation type and cover density. In areas where NOVI saturation did not occur, r2 values were best during peak greenness. When saturation was present, stronger relationships were evident after senescence of the dominant vegetation. Viewing geometry was also studied as a possible reason for differences in NDVI values between the three sensors. Pixels from the SPOT VEGETATION NDVI product were more frequently obtained in highly off nadir positions than those of AVHRR and MODIS. MODIS images had the fewest numbers of off nadir pixels. It is possible that efforts to minimize the effects of highly off nadir viewing may hinder the selection of the greenest pixels in Constrained View - Maximum Value Compositing.

Library of Congress Subject Headings

Vegetation mapping -- Remote sensing Detectors -- Evaluation



Number of Pages



South Dakota State University